- Email: [email protected]
Association between postoperative complications and lingering post-surgical pain: an observational cohort study
British Journal of Anaesthesia, xxx (xxx): xxx (xxxx) doi: 10.1016/j.bja.2019.10.012 Advance Access Publication Date: xxx Clinical Investigation
CLINICAL INVESTIGATION
Association between postoperative complications and lingering post-surgical pain: an observational cohort study Mark Willingham1, Govind Rangrass2, Caitlin Curcuru2, Arbi Ben Abdallah1, Troy S. Wildes1, Sherry McKinnon1, Alex Kronzer1, Anshuman Sharma1, Dan Helsten1, Bruce Hall3,4,5, Michael S. Avidan1 and Simon Haroutounian1,* 1
Department of Anesthesiology, Washington University in St Louis School of Medicine, St Louis, MO, USA, 2Department
of Anesthesia and Critical Care, University of Chicago Pritzker School of Medicine, Chicago, IL, USA, 3American College of Surgeons National Surgical Quality Improvement Program, Chicago IL, USA, 4Department of Surgery, Washington University in St Louis, St Louis, MO, USA and 5BJC HealthCare, St Louis, MO, USA *Corresponding author. E-mail: [email protected]
Abstract Background: Post-surgical pain that lingers beyond the initial few-week period of tissue healing is a major predictor of pain chronification, which leads to substantial disability and new persistent opioid analgesic use. We investigated whether postoperative medical complications increase the risk of lingering post-surgical pain. Methods: The study population consisted of patients undergoing diverse elective surgical procedures in an academic referral centre in the USA, between September 2013 and May 2017. Multivariable logistic regression, adjusting for confounding variables and patient-specific risk factors, was used to test for an independent association between any major postoperative complication and functionally limiting lingering pain 1e3 months after surgery, as obtained from patient self-reports. Results: The cohort included 11 986 adult surgical patients; 10 562 with complete data. At least one complication (cardiovascular, respiratory, renal/gastrointestinal, wound, thrombotic, or neural) was reported by 13.3% (95% confidence interval: 12.7e14.0) of patients, and 19.7% (19.0e20.5%) reported functionally limiting lingering post-surgical pain. After adjusting for known risk factors, the patients were twice as likely (odds ratio: 2.04; 1.78e2.35) to report lingering postsurgical pain if they also self-reported a postoperative complication. Experiencing a complication was also independently predictive of lingering post-surgical pain (odds ratio: 1.95; 1.26e3.04) when complication data were extracted from the National Surgical Quality Improvement Program registry, instead of being obtained from patient self-report. Conclusions: Medical complications were associated with a two-fold increase in functionally limiting pain 1e3 months after surgery. Understanding the mechanisms that link complications to pathological persistence of pain could help develop future approaches to prevent persistent post-surgical pain. Keywords: chronic post-surgical pain; lingering post-surgical pain; medical complications; pain; postoperative complications
Accepted: 12 October 2019 © 2019 British Journal of Anaesthesia. Published by Elsevier Ltd. All rights reserved. For Permissions, please email: [email protected]
1
2
-
Willingham et al.
Editor’s key points The role of postoperative complications in the development of persistent post-surgical pain is unclear. Using a large single-centre database with a questionnaire follow-up at least 30 days after surgery, at least one postoperative complication was reported by 13.3% of patients. After adjusting for other known risk factors for persistent pain, postoperative complications were associated with an increased risk or prolonged pain. These findings could be used to improve early diagnosis and management of persistent post-surgical pain through identifying high-risk individuals.
Persistent post-surgical pain (PPSP), defined as pain lasting more than 3e6 months after surgery, affects 10e50% of patients after common surgeries.1e5 Patients who report PPSP are often slow to recover and require more healthcare resources.6,7 In 10e20% of patients attending chronic pain clinics in the UK, the pain is attributable to a previous surgery.8 Furthermore, those with PPSP can become dependent on opioids, as evidenced by an estimated 2 million Americans transitioning to long-term opioid use annually after elective surgery,2,9,10 and 3e13.5% of surgical patients in Canadian cohorts reporting opioid use for more than 3 months after surgery.2,11,12 From a patient’s perspective, PPSP is debilitating and worsens the quality of life. In general, our understanding of why some patients develop PPSP whilst others do not remains limited. The type of surgery, surgical technique, and magnitude of tissue and nerve injury have all been implicated.13 Patient-related risk factors include female sex; younger age; obesity; and preexisting pain, anxiety, and depression.3,6,7 However, none of these factors have been found to be predictive across diverse surgeries. One of the major indicators of poor long-term post-surgical pain outcomes has been uncontrolled pain that lingers in the first weeks to months after surgery.14,15 This period may be critical for providing early interventions to prevent pain chronification16e18; therefore, understanding which factors contribute to pain that lingers beyond the acute postoperative period may inform PPSP prevention strategies. Basic science research suggests that dysregulation of the normal immune response and sustained inflammation may worsen acute pain and promote pain chronification.13 Ongoing release of pro-inflammatory mediators facilitates transition from acute to chronic pain,14e16 purportedly by hyperactivation of immune cells, such as microglia in the spinal cord, which contributes to sensitisation of the CNS and amplification of pain responses.19e21 Inflammatory conditions frequently accompany major medical complications, which occur in approximately 20% of surgical patients.22e27 Moreover, some data suggest that surgical-wound-related complications and surgery-related infections can increase the risk of PPSP.28,29 We, therefore, hypothesised that major medical complications, possibly by
exacerbating and prolonging the inflammatory response after surgery, would be associated with an increase in pain that lingers beyond the acute post-surgical period.
Methods This was an observational cohort study of unselected adult patients (18 yr) who underwent elective surgery at BarnesJewish Hospital (St Louis, MO, USA) between 1 September, 2013 and 31 May, 2017, and completed baseline and postoperative follow-up surveys as part of the Systematic Assessment and Targeted Improvement of Services Following Yearlong Surgical Outcomes Surveys (SATISFY-SOS; NCT02032030).30 All the analyses conducted in this study were pre-specified (Supplementary Appendix 1), and the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies were followed.31 The purpose of SATISFY-SOS was to track surgical patients’ postoperative health and well-being.30 Patients undergoing elective surgeries were enrolled at their preoperative visit, where they completed a baseline survey. Starting approximately 30 days after their surgery, the patients were contacted by e-mail to complete a single follow-up survey. Nonresponders were provided further opportunities to complete the survey via postal mail and telephone calls. All SATISFYSOS participants who returned the postoperative surveys and underwent invasive surgery were included in this study. Patients who were non-English speaking, not able to provide informed consent, under 18 yr of age, or suffering from dementia were excluded from SATISFY-SOS. This project was approved by the Washington University Institutional Review Board (IRB ID# 201605098). A waiver of informed consent for this study was obtained because the patients had already provided written informed consent to participate in SATISFYSOS. The primary outcome of this study was obtained from the SATISFY-SOS follow-up survey question: ‘During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)?’ The answers ‘quite a bit’ and ‘extremely’ were defined as severe lingering post-surgical pain, whilst the answers ‘moderately’, ‘a little bit’, and ‘not at all’ were defined as non-severe pain. Similarly, severe pre-existing pain was classified using an identical question on the baseline survey. The postoperative surveys also asked patients whether or not they experienced any of 17 major complications within 30 days of their surgery.32 These complications included (i) cardiac (myocardial infarction, cardiac arrest, congestive heart failure, and atrial fibrillation), (ii) respiratory (respiratory arrest, respiratory failure, and pneumonia), (iii) wound/infection (surgical wound infection and sepsis), (iv) thrombotic (deep vein thrombosis, pulmonary embolism, and stroke), (v) renal/gastrointestinal (dialysis, gastrointestinal bleeding, and gastrointestinal ulcer), and (vi) neural (nerve injury and paralysis). Complications were also gathered from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), a national registry for which dedicated, trained, and examined data collectors accrue rigorously specified surgical patient and case information, including 30-day postoperative outcomes.33e35 ACS-NSQIP complications spanned the same categories, except for neural complications, which were not
Postoperative complications and lingering pain
40 170 Potentially eligible patients visited the centre for preoperative assessment and planning
25 637 Consented to SATISFY-SOS
11 986 Completed baseline and postoperative surveys Primary analysis
Sensitivity analyses
1424 Excluded Non-exclusive missing categories: 808: Mental component score 488: Postoperative pain status 174: BMI 22: ASA physical status 2: Sex
10 562 Patients remain
11 986 Included after imputation of missing data
1590 Patients with complications reported by the National Surgical Quality Improvement Program
Fig 1. Flow diagram of cohort selection. SATISFY-SOS, Systematic Assessment and Targeted Improvement of Services Following Yearlong Surgical Outcomes Surveys.
available. These complications were used in a sensitivity analysis for the period under investigation. Potentially confounding factors, including patient characteristics, co-morbidities, baseline functionally limiting pain, and baseline emotional health, were also collected from the electronic medical record and the baseline SATISFY-SOS survey. Emotional health was obtained at baseline from the mental component score (MCS) derived from Veterans RAND 12 questions. This score ranges from 0 to 100, where a lower score is associated with decreased emotional health status and is normalised so that a score of 50 represents the US population mean. The patients were grouped by their surgical procedure into categories, including cardiothoracic surgery; general abdominal surgery (not minimally invasive); general abdominal surgery (minimally invasive); ear, nose, and throat/ plastic surgery; neurosurgery; orthopaedic surgery; urological/ gynaecological surgery; and vascular surgery. Information from SATISFY-SOS was collected, updated, and stored in a locally hosted SQL Server database. Patient characteristic and clinical data were collected from our anaesthesia information management system, MetaVision (iMDsoft, Needham, MA, USA). All data management and statistical analysis were completed using Microsoft SQL Server 12.0 (Microsoft, Redmond, WA, USA), Python 3.6 (Python Software Foundation, Wilmington, DE, USA), and SAS® version 9.4 (SAS Institute Inc., Cary, NC, USA).
Statistical analysis The patients were first stratified by surgical specialty and by whether they reported a major complication within 30 days of surgery. Within each category, we reported the proportion of patients who reported having severe pain over the past 4 weeks before their follow-up surveys. For patients experiencing any or no complication across all surgical categories, differences in baseline patient characteristics were compared with t-tests, WilcoxoneManneWhitney tests, and c2 tests.
-
3
Normality of continuous variables was assessed visually with histograms and by KolmogoroveSmirnov tests. A single-step multivariable logistic regression adjusting for confounding variables and patient-specific risk factors was performed to determine whether an independent association between the occurrence of any complication and lingering post-surgical pain exists. The following co-variables were preselected based on clinical supposition that they may be confounding factors: age, sex, BMI, pre-existing severe pain, ASA physical status score, and MCS. We also adjusted for duration of follow-up, as there may be significant differences in recovery throughout the post-surgical time period studied, for preoperative opioid use, and for surgical specialty. Age could not be entered as a continuous variable because it violated the assumption of linearity of the logit, so it was trichotomised into a reference group (<50 yr), middle age (50e64 yr), and older adults (65 yr). Additional logistic regressions with identical co-variables were performed to evaluate the associations between post-surgical pain and categories of complication types and within specific surgical specialties. For these regressions, 99.375% confidence intervals (CIs) were reported to approximate a Bonferroni-style adjustment for multiple comparisons. These CIs were created by setting alpha to 0.05/8, or 0.00625, to account for eight nearly identical regressions being performed. Two sensitivity analyses were performed. To evaluate for the impact of missing data, we performed multiple imputation for variables with <10% missing data (sex: two; ASA: 22; BMI: 178; lingering post-surgical pain: 488; and mental component score: 806). One hundred imputations were performed, and the results were combined with the SAS procedure MIANALYZE. To evaluate for the potential confounding impact of patient-reported complications, we conducted a subgroup analysis of surgical registry-reported complications from the ACS-NSQIP database in place of patient-reported complications. ACS-NSQIP complications were available from a random sampling of the same patient cohort who underwent surgery between November 12, 2013 and May 31, 2017, except for patients undergoing neurosurgical, orthopaedic, and head/neck surgeries.
Results From September 1, 2013 through May 31, 2017, 25 637 patients were enrolled in SATISFY-SOS and underwent an eligible procedure at Barnes-Jewish Hospital in St Louis, MO, USA. Of these, 11 986 completed the baseline and postoperative surveys (median response time: 68 days; inter-quartile range: 53e84 days; Fig. 1; Supplementary Fig. 1). Complete information was available for 10 562 individuals with a mean age of 59.2 (standard deviation: 13.9) yr, 58.8% of whom were female. Patient race was reported by 9091 participants, of whom 89.1% were Caucasian. The participants’ characteristic and surgical classifications were generally similar to patients who consented but did not participate, and to the overall patient population seen in the preoperative clinic (Table 1). Overall, one or more complications were reported by 13.3% of patients (95% CI: 12.7e14.0; Table 1). The most common type of postoperative complication was wound/infection (3.6%), followed by respiratory (3.0%), neural (2.8%), cardiac (2.5%), thrombotic (2.5%), and renal/gastrointestinal (2.0%). ACS-NSQIP data were available for 1590 patients, with a complication rate of 8.8% (95% CI: 7.4e10.2). Patients with severe pre-existing pain experienced an overall complication rate of 15.9% vs 12.4% in
4
-
Willingham et al.
Table 1 Characteristics of all eligible, consented, and included patients, and those who reported a complication. Data are presented as n (%), mean (standard deviation), or median (first to third quartiles), as appropriate. *Number with complete age, BMI, sex, ASA physical status, and surgery category. Percentage values in the ‘Complication’ column represent the incidence of complications in that row. ASA, American Society of Anesthesiologists; ENT, ear, nose, and throat; MCS, mental component score; MIS, minimally invasive (general) surgery; PCS, physical component score; VR-12, Veterans RAND 12-question survey
Age <50 yr Age 50e64 yr Age 65 yr BMI (kg m2) Female sex ASA physical status 1 2 3 4þ Severe baseline pain Baseline VR-12 MCS score Baseline VR-12 PCS score Preoperative opioid use Surgery category Open abdominal Cardiothoracic ENT/plastic MIS Neurosurgery Orthopaedic Urology/gynaecology Vascular surgery
Eligible (n¼31 831)*
Consented (n¼19 712)*
Study sample (n¼10 562)
Complication (n¼1408)
9782 12 243 9806 28.7 (24.7e33.8) 17 935
6137 7875 5700 29.1 (25.0e34.3) 11 433
2315 4316 3931 29.0 (25.1e34.0) 6214
297 (12.8) 595 (13.8) 516 (13.1) 29.3 (25.2e34.3) 799 (12.9)
1948 16 324 12 545 1014 d d d d
1255 10 530 7396 531 d d d d
588 5629 4018 327 2819 53.3 (11.1) 38.9 (12.4) 2564
52 (8.7) 574 (10.2) 664 (16.5) 119 (36.4) 448 (15.9) 50.7 (12.3) 36.3 (12.4) 372 (14.5)
5075 2191 4412 2788 2343 8446 6584 1469
3134 1357 2674 1821 1312 5405 4085 879
1180 746 1280 1111 734 3109 1985 417
209 (17.7) 208 (27.9) 177 (13.8) 107 (9.6) 104 (14.2) 315 (10.1) 201 (10.1) 87 (20.9)
those without (P<0.001). A similar relationship was found amongst NSQIP-defined complications (12.7% vs 8.1%; P¼0.023). Overall, 19.7% (95% CI: 19.0e20.5) of the cohort and 17.0% (95% CI: 15.1e18.8) of the subgroup, who were also in the ACSNSQIP registry, reported functionally limiting lingering postsurgical pain. In general, higher rates of complications were associated with more functional limitations from patients’ lingering postoperative pain (Supplementary Tables 1e7). Other unadjusted risk factors for reported lingering pain included age <65 yr (22.3% vs 15.5%; P<0.001), female sex (20.9% vs 18.1%; P<0.001), obesity (21.7% vs 18.2%; P<0.001), presence of pre-existing pain (42.6% vs 11.4%; P<0.001), normalised MCSs below 50 (31.1% vs 14.5%; P<0.001), preoperative opioid use (30.4% vs 16.3%; P<0.001), and ASA physical status scores 3 (23.6% vs 17.0%; P<0.001). After adjusting for these variables, the patients were twice as likely to report lingering pain if they also experienced a postoperative complication (Table 2; c-statistic¼0.77). This finding was consistent in a larger imputed data set and in patients for whom information on complications was obtained from the NSQIP registry (Table 2). Patients who returned their surveys later after their surgeries reported slightly less lingering post-surgical pain than those who returned their surveys earlier (Table 2). Our analysis demonstrated that a significant association between complications and lingering pain exists across multiple surgical specialties (adjusted odds ratios: 1.54e3.06; Fig. 2). Finally, after adjusting for confounding factors, each complication was associated with an increased risk of lingering pain, with odds ratios (95% CI) between 1.44 (1.06e1.96) for cardiovascular complications and 2.29 (1.74e3.01) for neural complications (Fig. 3).
Discussion This study has two findings with important clinical implications. First, it confirms that across diverse surgeries, including minimally invasive surgeries, 10e50% of patients reported lingering, functionally limiting pain at a median follow-up of 68 days. Second, major medical complications after elective surgery are associated with about a two-fold higher risk of developing such pain. Notably, apart from severe pre-existing pain, postoperative medical complications were the strongest independent predictors of lingering pain 1e3 months after surgery. These findings are in line with reports that surgical wound complications can increase the risk of persistent pain after cardiac surgery, and surgery-related infections increase the risk of persistent pain after hysterectomy.28,29 The degree to which patient or surgical factors drive some patients to have worse pain than others has been the subject of considerable debate. Whilst recognising that patientspecific factors, such as poor baseline mental health, may hinder recovery from pain, we must remain mindful that perioperative events may pose important ramifications on long-term pain outcomes. Similar to previous studies on long-term pain after surgery, we identified patient-specific risk factors, such as female sex, younger age, severe baseline pain, and poor baseline mental health.3,6,7 The strong association we identified between medical complications and lingering post-surgical pain is consistent with emerging research showing that patientspecific immune states are likely to be superior predictors of recovery from postoperative pain than other risk factors.36 Mounting evidence suggests that immune dysregulation, uncontrolled cytokine release, and a sustained inflammatory state may underpin the development of altered nociception
Postoperative complications and lingering pain
-
5
Table 2 Multivariable odds ratios for functionally limiting lingering post-surgical pain. ASA, American Society of Anesthesiologists; CI, confidence interval; ENT, ear, nose, and throat surgery; MI, multiple imputation of missing values; MIS, minimally invasive surgery; NSQIP, National Surgical Quality Improvement Program; SATISFY-SOS, Systematic Assessment and Targeted Improvement of Services Following Yearlong Surgical Outcomes Surveys Factor
Any complication Severe baseline pain Age 65 (vs <50) yr Age 50e64 (vs <50) yr Female sex (vs male) BMI (per kg m2) ASA physical status (vs 1) 2 3 4 Mental component score (per point) Preoperative opioid use Time to follow-up survey (per day) Surgical service (vs MIS) Orthopaedic Cardiothoracic Vascular Abdominal Urology/gynaecology ENT/plastic Neurosurgery
SATISFY-SOS odds ratio (95% CI)
SATISFY-SOS MI odds ratio (95% CI)
Conditional NSQIP odds ratio (95% CI)
(n¼10 562)
P-value
(n¼11 986)
P-value
(n¼1590)
P-value
2.04 4.02 0.62 0.99 1.15 1.00
(1.78e2.35) (3.58e4.53) (0.53e0.73) (0.86e1.14) (1.03e1.28) (1.00e1.01)
<0.001 <0.001 <0.001 0.901 0.013 0.313
2.03 4.06 0.61 1.01 1.12 1.00
(1.78e2.32) (3.63e4.54) (0.53e0.71) (0.88e1.15) (1.01e1.25) (1.00e1.01)
<0.001 <0.001 <0.001 0.887 0.033 0.147
1.95 3.48 0.65 1.08 1.40 1.00
(1.26e3.04) (2.46e4.92) (0.42e1.02) (0.73e1.59) (1.02e1.91) (0.98e1.02)
0.003 <0.001 0.06 0.708 0.037 0.884
1.13 1.65 1.29 0.97 1.42 1.00
(0.88e1.45) (1.27e2.15) (0.85e1.97) (0.97e0.98) (1.27e1.60) (0.99e1.00)
0.296 <0.001 0.293 <0.001 <0.001 <0.001
1.10 1.56 1.22 0.97 1.43 0.99
(0.86e1.40) (1.21e2.01) (0.82e1.82) (0.97e0.98) (1.28e1.60) (0.99e1.00)
0.444 <0.001 0.327 <0.001 <0.001 <0.001
0.60 0.75 0.36 0.98 2.49 0.99
(0.24e1.51) (0.29e1.93) (0.09e1.39) (0.96e0.99) (1.82e3.42) (0.99e1.00)
0.279 0.552 0.138 <0.001 <0.001 0.057
1.61 1.16 1.54 1.24 0.98 1.03 1.34
(1.30e1.99) (0.86e1.55) (1.13e2.11) (0.97e1.59) (0.78e1.23) (0.80e1.32) (1.03e1.74)
<0.001 0.313 0.006 0.089 0.944 0.893 0.021
1.64 1.27 1.59 1.24 1.02 1.07 1.36
(1.34e2.01) (0.96e1.68) (1.18e2.15) (0.98e1.58) (0.82e1.27) (0.84e1.36) (1.06e1.74)
<0.001 0.093 0.002 0.078 0.857 0.564 0.017
1.69 1.49 1.49 1.15 0.93
d (0.92e3.11) (0.86e2.58) (0.96e2.3) (0.73e1.80) (0.44e1.98) d
d 0.092 0.153 0.073 0.555 0.849 d
and hyperalgesia.37,38 In the perioperative setting, medical complications may further exacerbate this inflammation, leading to persistent pain after surgery. The finding that no single complication appeared to be disproportionately responsible for the development of lingering pain could imply a systemic underlying mechanism, such as inflammatory or immune response, rather than a locally mediated effect. However, an association between severe acute pain on postoperative Day 1 and 30-day postoperative complications has been reported, questioning the directionality of the association.39 Therefore, future studies are required for understanding the precise mechanisms linking postoperative complications with persistent pain, and for finding biomarkers that can help identify high-risk patients in this context. This study has notable strengths. The data were collected prospectively using validated survey instruments from a large number of patients who underwent diverse elective surgical procedures. As such, there is a strong probability that these findings are generalisable. The credibility of the main finding is bolstered by the sensitivity analyses, including the incorporation of complications from the rigorous and established ACS-NSQIP registry. Furthermore, the association between medical complications and lingering post-surgical pain is consistent across different surgeries and medical complications affecting multiple organ systems, and after controlling for variables, such as age, sex, and the presence of pre-surgical pain. Similarly, the study has important limitations. First, with an observational design and retrospective analysis, causality cannot be inferred. Our findings should therefore be interpreted as hypothesis generating. Our pre-specified study protocol (Supplementary material) also slightly overestimated the number of available patients, because many entries in the initially obtained data set were for patients who underwent
anaesthesia for an excluded procedure (e.g. percutaneous intervention, colonoscopy, electroconvulsive therapy, etc.) or underwent numerous procedures. However, we were still able to include pain outcomes from a large and diverse surgical cohort. Second, our primary analysis utilised patient selfreport to assess complications, which may be subject to recall bias and inaccuracy because of patient misinterpretation, survey fatigue, or temporal ambiguity regarding perioperative events. A previous study using our patient-reported outcomes (PRO) method showed an excellent negative agreement, but only a low-to-moderate positive agreement in the detection of complications, when comparing PROs with complications documented in the electronic medical record.34 Whilst this finding might have revealed limitations in the PROs, it could also have reflected an advantage in that, with SATISFY-SOS, the patients reported complications that occurred both during hospitalisation and after discharge, whereas post-discharge complications are often not captured in the electronic medical record. We also considered that patients might be unsure about their medical diagnoses or complications, especially if they had more complicated hospital courses. To address this potential limitation, we conducted a sensitivity analysis using well-validated NSQIP complication data for patients enrolled in SATISFY-SOS, and reached similar findings. Third, the time frame of lingering pain outcome assessment in our study was generally within 1e3 months after surgery. Whilst it is highly probable that these outcomes are highly linked with the development of PPSP, the proposed cut-offs for which have ranged from 2 to 6 months after surgery,1 additional research is required for understanding the association between postoperative complications and outcomes beyond the 6-month postoperative time frame. Finally, our outcome of interest did not specify surgical site pain. The lack of pain location is somewhat mitigated by
6
-
Willingham et al.
Fig 2. Proportion with lingering post-surgical pain, and adjusted odds for lingering pain if a complication was also reported. Confidence intervals (CI) are Bonferroni corrected for eight regressions and include adjustment for baseline pain, age, sex, BMI, ASA physical status, preoperative opioid use, and mental component score. ENT, ear, nose, and throat; MIS, minimally invasive surgery; OR, odds ratio.
our inclusion of preoperative baseline pain (asked as exactly the same question) in our regression analysis. Of note, the non-specificity of this outcome may be appropriate, as inflammation from surgery and postoperative complications might globally sensitise patients to pain, including subsidiary pain from existing co-morbidities. In conclusion, this study adds to existing evidence by confirming in a large and diverse surgical cohort that lingering
post-surgical pain is common. Most importantly, our hypothesis that patients who experience postoperative medical complications are likely to be at much higher risk for pain that persists 1e3 months after surgery was supported. This finding, if replicated in future studies, suggests that clinicians should target patients who experience such complications for candidate interventions that might be effective in preventing postsurgical pain chronification. As long-term pain after surgery
Fig 3. Independent associations between complication categories and lingering post-surgical pain. Odds ratios are adjusted for baseline pain, age, sex, BMI, ASA physical status, mental component score, preoperative opioid use, and duration of follow-up. Odds ratios (95% confidence interval) from top to bottom are 1.44 (1.06e1.96), 1.53 (1.18e1.98), 1.56 (1.18e2.05), 1.60 (1.15e2.25), 1.66 (1.23e2.23), and 2.29 (1.74e3.01). GI, gastrointestinal.
Postoperative complications and lingering pain
is emerging as a major source of disability and persistent opioid use,2,9 these findings may have important implications for preventing these adverse consequences of surgery in highrisk patients.
Authors’ contributions Study conception/design: all authors. Funds acquisition: MSA. Data acquisition/analysis/interpretation: MW, GR, CC, ABA, AK, DH, MSA, SH. Statistical analyses: MW, GR. Drafting of manuscript: MW, GR. SH had full access to all the data in the study, and takes responsibility for the integrity of the data and the accuracy of the data analysis. He is the guarantor. All authors critically revised the manuscript for important intellectual content, and have approved the final version of the manuscript.
7.
8. 9.
10.
11.
12.
Acknowledgements The funding sources provided infrastructure and financial support, but had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication.
Declarations of interest
13.
14.
15.
The authors declare that they have no competing interests related to the publication or content of this publication. MSA is an editor of the British Journal of Anaesthesia. 16.
Funding Foundation for Barnes-Jewish Hospital (7937-77); Medical Student Anesthesia Research Fellowship grant from the Foundation for Anesthesia Education and Research to CC.
17.
18.
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.bja.2019.10.012.
19.
References 1. Werner MU, Kongsgaard UEI. Defining persistent postsurgical pain: is an update required? Br J Anaesth 2014; 113: 1e4 2. Clarke H, Soneji N, Ko DT, Yun L, Wijeysundera DN. Rates and risk factors for prolonged opioid use after major surgery: population based cohort study. BMJ 2014; 348: g1251 3. Kehlet H, Jensen TS, Woolf CJ. Persistent postsurgical pain: risk factors and prevention. Lancet 2006; 367: 1618e25 4. Kehlet H, Rathmell JP. Persistent postsurgical pain: the path forward through better design of clinical studies. Anesthesiology 2010; 112: 514e5 5. Haroutounian S, Nikolajsen L, Bendtsen TF, et al. Primary afferent input critical for maintaining spontaneous pain in peripheral neuropathy. Pain 2014; 155: 1272e9 6. Parsons B, Schaefer C, Mann R, et al. Economic and humanistic burden of post-trauma and post-surgical
20.
21.
22.
23.
-
7
neuropathic pain among adults in the United States. J Pain Res 2013; 6: 459e69 VanDenKerkhof EG, Hopman WM, Reitsma ML, et al. Chronic pain, healthcare utilization, and quality of life following gastrointestinal surgery. Can J Anaesth 2012; 59: 670e80 Macrae WA. Chronic pain after surgery. Br J Anaesth 2001; 87: 88e98 Brummett CM, Waljee JF, Goesling J, et al. New persistent opioid use after minor and major surgical procedures in US adults. JAMA Surg 2017; 152, e170504 Sun EC, Darnall BD, Baker LC, Mackey S. Incidence of and risk factors for chronic opioid use among opioid-naive patients in the postoperative period. JAMA Intern Med 2016; 176: 1286e93 Huang A, Azam A, Segal S, et al. Chronic postsurgical pain and persistent opioid use following surgery: the need for a transitional pain service. Pain Manag 2016; 6: 435e43 Soneji N, Clarke HA, Ko DT, Wijeysundera DN. Risks of developing persistent opioid use after major surgery. JAMA Surg 2016; 151: 1083e4 Wildgaard K, Ravn J, Kehlet H. Chronic post-thoracotomy pain: a critical review of pathogenic mechanisms and strategies for prevention. Eur J Cardiothorac Surg 2009; 36: 170e80 Aasvang EK, Luna IE, Kehlet H. Challenges in postdischarge function and recovery: the case of fast-track hip and knee arthroplasty. Br J Anaesth 2015; 115: 861e6 Clarke H. Transitional pain medicine: novel pharmacological treatments for the management of moderate to severe postsurgical pain. Expert Rev Clin Pharmacol 2016; 9: 345e9 Katz J, Jackson M, Kavanagh BP, Sandler AN. Acute pain after thoracic surgery predicts long-term post-thoracotomy pain. Clin J Pain 1996; 12: 50e5 Katz J, Seltzer Z. Transition from acute to chronic postsurgical pain: risk factors and protective factors. Expert Rev Neurother 2009; 9: 723e44 Katz J, Weinrib A, Fashler SR, et al. The Toronto General Hospital Transitional Pain Service: development and implementation of a multidisciplinary program to prevent chronic postsurgical pain. J Pain Res 2015; 8: 695e702 Nazemi S, Manaheji H, Noorbakhsh SM, et al. Inhibition of microglial activity alters spinal wide dynamic range neuron discharge and reduces microglial Toll-like receptor 4 expression in neuropathic rats. Clin Exp Pharmacol Physiol 2015; 42: 772e9 Yamamoto Y, Terayama R, Kishimoto N, et al. Activated microglia contribute to convergent nociceptive inputs to spinal dorsal horn neurons and the development of neuropathic pain. Neurochem Res 2015; 40: 1000e12 Zhou L-J, Yang T, Wei X, et al. Brain-derived neurotrophic factor contributes to spinal long-term potentiation and mechanical hypersensitivity by activation of spinal microglia in rat. Brain Behav Immun 2011; 25: 322e34 Duncan DJ, Yang Z, Hopkins PM, Steele DS, Harrison SM. TNF-alpha and IL-1beta increase Ca2þ leak from the sarcoplasmic reticulum and susceptibility to arrhythmia in rat ventricular myocytes. Cell Calcium 2010; 47: 378e86 Hotchkiss RS, Monneret G, Payen D. Immunosuppression in sepsis: a novel understanding of the disorder and a new therapeutic approach. Lancet Infect Dis 2013; 13: 260e8
8
-
Willingham et al.
24. Liu YY, Chiang CH, Chuang CH, Liu SL, Jheng YH, Ryu JH. Spillover of cytokines and reactive oxygen species in ventilator-induced lung injury associated with inflammation and apoptosis in distal organs. Respir Care 2014; 59: 1422e32 25. Singbartl K, Joannidis M. Short-term effects of acute kidney injury. Crit Care Clin 2015; 31: 751e62 26. Christiansen SC, Naess IA, Cannegieter SC, Hammerstrom J, Rosendaal FR, Reitsma PH. Inflammatory cytokines as risk factors for a first venous thrombosis: a prospective population-based study. PLoS Med 2006; 3: e334 27. Ghaferi AA, Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery. N Engl J Med 2009; 361: 1368e75 28. Costa MA, Trentini CA, Schafranski MD, Pipino O, Gomes RZ, Reis ES. Factors associated with the development of chronic post-sternotomy pain: a case-control study. Braz J Cardiovasc Surg 2015; 30: 552e6 29. Theunissen M, Peters ML, Schepers J, et al. Recovery 3 and 12 months after hysterectomy: epidemiology and predictors of chronic pain, physical functioning, and global surgical recovery. Medicine (Baltimore) 2016; 95, e3980 30. Helsten DL, Ben Abdallah A, Avidan MS, et al. Methodologic considerations for collecting patient-reported outcomes from unselected surgical patients. Anesthesiology 2016; 125: 495e504 31. von Elm E, Altman DG, Egger M, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ 2007; 335: 806e8
32. Fritz BA, Escallier KE, Ben Abdallah A, et al. Convergent validity of three methods for measuring postoperative complications. Anesthesiology 2016; 124: 1265e76 33. Hall BL, Hamilton BH, Richards K, Bilimoria KY, Cohen ME, Ko CY. Does surgical quality improve in the American College of Surgeons national surgical quality improvement program: an evaluation of all participating hospitals. Ann Surg 2009; 250: 363e76 34. Cohen ME, Liu Y, Ko CY, Hall BL. Improved surgical outcomes for ACS NSQIP hospitals over time: evaluation of hospital cohorts with up to 8 years of participation. Ann Surg 2016; 263: 267e73 35. Khuri SF. The NSQIP: a new frontier in surgery. Surgery 2005; 138: 837e43 36. Fragiadakis GK, Gaudilliere B, Ganio EA, et al. Patientspecific immune states before surgery are strong correlates of surgical recovery. Anesthesiology 2015; 123: 1241e55 37. Rowe RK, Ellis GI, Harrison JL, et al. Diffuse traumatic brain injury induces prolonged immune dysregulation and potentiates hyperalgesia following a peripheral immune challenge. Mol Pain 2016; 12: 1e12 38. Bugada D, Lavand’homme P, Ambrosoli AL, et al. Effect of preoperative inflammatory status and comorbidities on pain resolution and persistent postsurgical pain after inguinal hernia repair. Mediators Inflamm 2016; 2016: 5830347 39. van Boekel RLM, Warle MC, Nielen RGC, et al. Relationship between postoperative pain and overall 30-day complications in a broad surgical population: an observational study. Ann Surg 2019; 269: 856e65 Handling editor: Lesley Colvin
CLINICAL INVESTIGATION
Association between postoperative complications and lingering post-surgical pain: an observational cohort study Mark Willingham1, Govind Rangrass2, Caitlin Curcuru2, Arbi Ben Abdallah1, Troy S. Wildes1, Sherry McKinnon1, Alex Kronzer1, Anshuman Sharma1, Dan Helsten1, Bruce Hall3,4,5, Michael S. Avidan1 and Simon Haroutounian1,* 1
Department of Anesthesiology, Washington University in St Louis School of Medicine, St Louis, MO, USA, 2Department
of Anesthesia and Critical Care, University of Chicago Pritzker School of Medicine, Chicago, IL, USA, 3American College of Surgeons National Surgical Quality Improvement Program, Chicago IL, USA, 4Department of Surgery, Washington University in St Louis, St Louis, MO, USA and 5BJC HealthCare, St Louis, MO, USA *Corresponding author. E-mail: [email protected]
Abstract Background: Post-surgical pain that lingers beyond the initial few-week period of tissue healing is a major predictor of pain chronification, which leads to substantial disability and new persistent opioid analgesic use. We investigated whether postoperative medical complications increase the risk of lingering post-surgical pain. Methods: The study population consisted of patients undergoing diverse elective surgical procedures in an academic referral centre in the USA, between September 2013 and May 2017. Multivariable logistic regression, adjusting for confounding variables and patient-specific risk factors, was used to test for an independent association between any major postoperative complication and functionally limiting lingering pain 1e3 months after surgery, as obtained from patient self-reports. Results: The cohort included 11 986 adult surgical patients; 10 562 with complete data. At least one complication (cardiovascular, respiratory, renal/gastrointestinal, wound, thrombotic, or neural) was reported by 13.3% (95% confidence interval: 12.7e14.0) of patients, and 19.7% (19.0e20.5%) reported functionally limiting lingering post-surgical pain. After adjusting for known risk factors, the patients were twice as likely (odds ratio: 2.04; 1.78e2.35) to report lingering postsurgical pain if they also self-reported a postoperative complication. Experiencing a complication was also independently predictive of lingering post-surgical pain (odds ratio: 1.95; 1.26e3.04) when complication data were extracted from the National Surgical Quality Improvement Program registry, instead of being obtained from patient self-report. Conclusions: Medical complications were associated with a two-fold increase in functionally limiting pain 1e3 months after surgery. Understanding the mechanisms that link complications to pathological persistence of pain could help develop future approaches to prevent persistent post-surgical pain. Keywords: chronic post-surgical pain; lingering post-surgical pain; medical complications; pain; postoperative complications
Accepted: 12 October 2019 © 2019 British Journal of Anaesthesia. Published by Elsevier Ltd. All rights reserved. For Permissions, please email: [email protected]
1
2
-
Willingham et al.
Editor’s key points The role of postoperative complications in the development of persistent post-surgical pain is unclear. Using a large single-centre database with a questionnaire follow-up at least 30 days after surgery, at least one postoperative complication was reported by 13.3% of patients. After adjusting for other known risk factors for persistent pain, postoperative complications were associated with an increased risk or prolonged pain. These findings could be used to improve early diagnosis and management of persistent post-surgical pain through identifying high-risk individuals.
Persistent post-surgical pain (PPSP), defined as pain lasting more than 3e6 months after surgery, affects 10e50% of patients after common surgeries.1e5 Patients who report PPSP are often slow to recover and require more healthcare resources.6,7 In 10e20% of patients attending chronic pain clinics in the UK, the pain is attributable to a previous surgery.8 Furthermore, those with PPSP can become dependent on opioids, as evidenced by an estimated 2 million Americans transitioning to long-term opioid use annually after elective surgery,2,9,10 and 3e13.5% of surgical patients in Canadian cohorts reporting opioid use for more than 3 months after surgery.2,11,12 From a patient’s perspective, PPSP is debilitating and worsens the quality of life. In general, our understanding of why some patients develop PPSP whilst others do not remains limited. The type of surgery, surgical technique, and magnitude of tissue and nerve injury have all been implicated.13 Patient-related risk factors include female sex; younger age; obesity; and preexisting pain, anxiety, and depression.3,6,7 However, none of these factors have been found to be predictive across diverse surgeries. One of the major indicators of poor long-term post-surgical pain outcomes has been uncontrolled pain that lingers in the first weeks to months after surgery.14,15 This period may be critical for providing early interventions to prevent pain chronification16e18; therefore, understanding which factors contribute to pain that lingers beyond the acute postoperative period may inform PPSP prevention strategies. Basic science research suggests that dysregulation of the normal immune response and sustained inflammation may worsen acute pain and promote pain chronification.13 Ongoing release of pro-inflammatory mediators facilitates transition from acute to chronic pain,14e16 purportedly by hyperactivation of immune cells, such as microglia in the spinal cord, which contributes to sensitisation of the CNS and amplification of pain responses.19e21 Inflammatory conditions frequently accompany major medical complications, which occur in approximately 20% of surgical patients.22e27 Moreover, some data suggest that surgical-wound-related complications and surgery-related infections can increase the risk of PPSP.28,29 We, therefore, hypothesised that major medical complications, possibly by
exacerbating and prolonging the inflammatory response after surgery, would be associated with an increase in pain that lingers beyond the acute post-surgical period.
Methods This was an observational cohort study of unselected adult patients (18 yr) who underwent elective surgery at BarnesJewish Hospital (St Louis, MO, USA) between 1 September, 2013 and 31 May, 2017, and completed baseline and postoperative follow-up surveys as part of the Systematic Assessment and Targeted Improvement of Services Following Yearlong Surgical Outcomes Surveys (SATISFY-SOS; NCT02032030).30 All the analyses conducted in this study were pre-specified (Supplementary Appendix 1), and the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies were followed.31 The purpose of SATISFY-SOS was to track surgical patients’ postoperative health and well-being.30 Patients undergoing elective surgeries were enrolled at their preoperative visit, where they completed a baseline survey. Starting approximately 30 days after their surgery, the patients were contacted by e-mail to complete a single follow-up survey. Nonresponders were provided further opportunities to complete the survey via postal mail and telephone calls. All SATISFYSOS participants who returned the postoperative surveys and underwent invasive surgery were included in this study. Patients who were non-English speaking, not able to provide informed consent, under 18 yr of age, or suffering from dementia were excluded from SATISFY-SOS. This project was approved by the Washington University Institutional Review Board (IRB ID# 201605098). A waiver of informed consent for this study was obtained because the patients had already provided written informed consent to participate in SATISFYSOS. The primary outcome of this study was obtained from the SATISFY-SOS follow-up survey question: ‘During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)?’ The answers ‘quite a bit’ and ‘extremely’ were defined as severe lingering post-surgical pain, whilst the answers ‘moderately’, ‘a little bit’, and ‘not at all’ were defined as non-severe pain. Similarly, severe pre-existing pain was classified using an identical question on the baseline survey. The postoperative surveys also asked patients whether or not they experienced any of 17 major complications within 30 days of their surgery.32 These complications included (i) cardiac (myocardial infarction, cardiac arrest, congestive heart failure, and atrial fibrillation), (ii) respiratory (respiratory arrest, respiratory failure, and pneumonia), (iii) wound/infection (surgical wound infection and sepsis), (iv) thrombotic (deep vein thrombosis, pulmonary embolism, and stroke), (v) renal/gastrointestinal (dialysis, gastrointestinal bleeding, and gastrointestinal ulcer), and (vi) neural (nerve injury and paralysis). Complications were also gathered from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), a national registry for which dedicated, trained, and examined data collectors accrue rigorously specified surgical patient and case information, including 30-day postoperative outcomes.33e35 ACS-NSQIP complications spanned the same categories, except for neural complications, which were not
Postoperative complications and lingering pain
40 170 Potentially eligible patients visited the centre for preoperative assessment and planning
25 637 Consented to SATISFY-SOS
11 986 Completed baseline and postoperative surveys Primary analysis
Sensitivity analyses
1424 Excluded Non-exclusive missing categories: 808: Mental component score 488: Postoperative pain status 174: BMI 22: ASA physical status 2: Sex
10 562 Patients remain
11 986 Included after imputation of missing data
1590 Patients with complications reported by the National Surgical Quality Improvement Program
Fig 1. Flow diagram of cohort selection. SATISFY-SOS, Systematic Assessment and Targeted Improvement of Services Following Yearlong Surgical Outcomes Surveys.
available. These complications were used in a sensitivity analysis for the period under investigation. Potentially confounding factors, including patient characteristics, co-morbidities, baseline functionally limiting pain, and baseline emotional health, were also collected from the electronic medical record and the baseline SATISFY-SOS survey. Emotional health was obtained at baseline from the mental component score (MCS) derived from Veterans RAND 12 questions. This score ranges from 0 to 100, where a lower score is associated with decreased emotional health status and is normalised so that a score of 50 represents the US population mean. The patients were grouped by their surgical procedure into categories, including cardiothoracic surgery; general abdominal surgery (not minimally invasive); general abdominal surgery (minimally invasive); ear, nose, and throat/ plastic surgery; neurosurgery; orthopaedic surgery; urological/ gynaecological surgery; and vascular surgery. Information from SATISFY-SOS was collected, updated, and stored in a locally hosted SQL Server database. Patient characteristic and clinical data were collected from our anaesthesia information management system, MetaVision (iMDsoft, Needham, MA, USA). All data management and statistical analysis were completed using Microsoft SQL Server 12.0 (Microsoft, Redmond, WA, USA), Python 3.6 (Python Software Foundation, Wilmington, DE, USA), and SAS® version 9.4 (SAS Institute Inc., Cary, NC, USA).
Statistical analysis The patients were first stratified by surgical specialty and by whether they reported a major complication within 30 days of surgery. Within each category, we reported the proportion of patients who reported having severe pain over the past 4 weeks before their follow-up surveys. For patients experiencing any or no complication across all surgical categories, differences in baseline patient characteristics were compared with t-tests, WilcoxoneManneWhitney tests, and c2 tests.
-
3
Normality of continuous variables was assessed visually with histograms and by KolmogoroveSmirnov tests. A single-step multivariable logistic regression adjusting for confounding variables and patient-specific risk factors was performed to determine whether an independent association between the occurrence of any complication and lingering post-surgical pain exists. The following co-variables were preselected based on clinical supposition that they may be confounding factors: age, sex, BMI, pre-existing severe pain, ASA physical status score, and MCS. We also adjusted for duration of follow-up, as there may be significant differences in recovery throughout the post-surgical time period studied, for preoperative opioid use, and for surgical specialty. Age could not be entered as a continuous variable because it violated the assumption of linearity of the logit, so it was trichotomised into a reference group (<50 yr), middle age (50e64 yr), and older adults (65 yr). Additional logistic regressions with identical co-variables were performed to evaluate the associations between post-surgical pain and categories of complication types and within specific surgical specialties. For these regressions, 99.375% confidence intervals (CIs) were reported to approximate a Bonferroni-style adjustment for multiple comparisons. These CIs were created by setting alpha to 0.05/8, or 0.00625, to account for eight nearly identical regressions being performed. Two sensitivity analyses were performed. To evaluate for the impact of missing data, we performed multiple imputation for variables with <10% missing data (sex: two; ASA: 22; BMI: 178; lingering post-surgical pain: 488; and mental component score: 806). One hundred imputations were performed, and the results were combined with the SAS procedure MIANALYZE. To evaluate for the potential confounding impact of patient-reported complications, we conducted a subgroup analysis of surgical registry-reported complications from the ACS-NSQIP database in place of patient-reported complications. ACS-NSQIP complications were available from a random sampling of the same patient cohort who underwent surgery between November 12, 2013 and May 31, 2017, except for patients undergoing neurosurgical, orthopaedic, and head/neck surgeries.
Results From September 1, 2013 through May 31, 2017, 25 637 patients were enrolled in SATISFY-SOS and underwent an eligible procedure at Barnes-Jewish Hospital in St Louis, MO, USA. Of these, 11 986 completed the baseline and postoperative surveys (median response time: 68 days; inter-quartile range: 53e84 days; Fig. 1; Supplementary Fig. 1). Complete information was available for 10 562 individuals with a mean age of 59.2 (standard deviation: 13.9) yr, 58.8% of whom were female. Patient race was reported by 9091 participants, of whom 89.1% were Caucasian. The participants’ characteristic and surgical classifications were generally similar to patients who consented but did not participate, and to the overall patient population seen in the preoperative clinic (Table 1). Overall, one or more complications were reported by 13.3% of patients (95% CI: 12.7e14.0; Table 1). The most common type of postoperative complication was wound/infection (3.6%), followed by respiratory (3.0%), neural (2.8%), cardiac (2.5%), thrombotic (2.5%), and renal/gastrointestinal (2.0%). ACS-NSQIP data were available for 1590 patients, with a complication rate of 8.8% (95% CI: 7.4e10.2). Patients with severe pre-existing pain experienced an overall complication rate of 15.9% vs 12.4% in
4
-
Willingham et al.
Table 1 Characteristics of all eligible, consented, and included patients, and those who reported a complication. Data are presented as n (%), mean (standard deviation), or median (first to third quartiles), as appropriate. *Number with complete age, BMI, sex, ASA physical status, and surgery category. Percentage values in the ‘Complication’ column represent the incidence of complications in that row. ASA, American Society of Anesthesiologists; ENT, ear, nose, and throat; MCS, mental component score; MIS, minimally invasive (general) surgery; PCS, physical component score; VR-12, Veterans RAND 12-question survey
Age <50 yr Age 50e64 yr Age 65 yr BMI (kg m2) Female sex ASA physical status 1 2 3 4þ Severe baseline pain Baseline VR-12 MCS score Baseline VR-12 PCS score Preoperative opioid use Surgery category Open abdominal Cardiothoracic ENT/plastic MIS Neurosurgery Orthopaedic Urology/gynaecology Vascular surgery
Eligible (n¼31 831)*
Consented (n¼19 712)*
Study sample (n¼10 562)
Complication (n¼1408)
9782 12 243 9806 28.7 (24.7e33.8) 17 935
6137 7875 5700 29.1 (25.0e34.3) 11 433
2315 4316 3931 29.0 (25.1e34.0) 6214
297 (12.8) 595 (13.8) 516 (13.1) 29.3 (25.2e34.3) 799 (12.9)
1948 16 324 12 545 1014 d d d d
1255 10 530 7396 531 d d d d
588 5629 4018 327 2819 53.3 (11.1) 38.9 (12.4) 2564
52 (8.7) 574 (10.2) 664 (16.5) 119 (36.4) 448 (15.9) 50.7 (12.3) 36.3 (12.4) 372 (14.5)
5075 2191 4412 2788 2343 8446 6584 1469
3134 1357 2674 1821 1312 5405 4085 879
1180 746 1280 1111 734 3109 1985 417
209 (17.7) 208 (27.9) 177 (13.8) 107 (9.6) 104 (14.2) 315 (10.1) 201 (10.1) 87 (20.9)
those without (P<0.001). A similar relationship was found amongst NSQIP-defined complications (12.7% vs 8.1%; P¼0.023). Overall, 19.7% (95% CI: 19.0e20.5) of the cohort and 17.0% (95% CI: 15.1e18.8) of the subgroup, who were also in the ACSNSQIP registry, reported functionally limiting lingering postsurgical pain. In general, higher rates of complications were associated with more functional limitations from patients’ lingering postoperative pain (Supplementary Tables 1e7). Other unadjusted risk factors for reported lingering pain included age <65 yr (22.3% vs 15.5%; P<0.001), female sex (20.9% vs 18.1%; P<0.001), obesity (21.7% vs 18.2%; P<0.001), presence of pre-existing pain (42.6% vs 11.4%; P<0.001), normalised MCSs below 50 (31.1% vs 14.5%; P<0.001), preoperative opioid use (30.4% vs 16.3%; P<0.001), and ASA physical status scores 3 (23.6% vs 17.0%; P<0.001). After adjusting for these variables, the patients were twice as likely to report lingering pain if they also experienced a postoperative complication (Table 2; c-statistic¼0.77). This finding was consistent in a larger imputed data set and in patients for whom information on complications was obtained from the NSQIP registry (Table 2). Patients who returned their surveys later after their surgeries reported slightly less lingering post-surgical pain than those who returned their surveys earlier (Table 2). Our analysis demonstrated that a significant association between complications and lingering pain exists across multiple surgical specialties (adjusted odds ratios: 1.54e3.06; Fig. 2). Finally, after adjusting for confounding factors, each complication was associated with an increased risk of lingering pain, with odds ratios (95% CI) between 1.44 (1.06e1.96) for cardiovascular complications and 2.29 (1.74e3.01) for neural complications (Fig. 3).
Discussion This study has two findings with important clinical implications. First, it confirms that across diverse surgeries, including minimally invasive surgeries, 10e50% of patients reported lingering, functionally limiting pain at a median follow-up of 68 days. Second, major medical complications after elective surgery are associated with about a two-fold higher risk of developing such pain. Notably, apart from severe pre-existing pain, postoperative medical complications were the strongest independent predictors of lingering pain 1e3 months after surgery. These findings are in line with reports that surgical wound complications can increase the risk of persistent pain after cardiac surgery, and surgery-related infections increase the risk of persistent pain after hysterectomy.28,29 The degree to which patient or surgical factors drive some patients to have worse pain than others has been the subject of considerable debate. Whilst recognising that patientspecific factors, such as poor baseline mental health, may hinder recovery from pain, we must remain mindful that perioperative events may pose important ramifications on long-term pain outcomes. Similar to previous studies on long-term pain after surgery, we identified patient-specific risk factors, such as female sex, younger age, severe baseline pain, and poor baseline mental health.3,6,7 The strong association we identified between medical complications and lingering post-surgical pain is consistent with emerging research showing that patientspecific immune states are likely to be superior predictors of recovery from postoperative pain than other risk factors.36 Mounting evidence suggests that immune dysregulation, uncontrolled cytokine release, and a sustained inflammatory state may underpin the development of altered nociception
Postoperative complications and lingering pain
-
5
Table 2 Multivariable odds ratios for functionally limiting lingering post-surgical pain. ASA, American Society of Anesthesiologists; CI, confidence interval; ENT, ear, nose, and throat surgery; MI, multiple imputation of missing values; MIS, minimally invasive surgery; NSQIP, National Surgical Quality Improvement Program; SATISFY-SOS, Systematic Assessment and Targeted Improvement of Services Following Yearlong Surgical Outcomes Surveys Factor
Any complication Severe baseline pain Age 65 (vs <50) yr Age 50e64 (vs <50) yr Female sex (vs male) BMI (per kg m2) ASA physical status (vs 1) 2 3 4 Mental component score (per point) Preoperative opioid use Time to follow-up survey (per day) Surgical service (vs MIS) Orthopaedic Cardiothoracic Vascular Abdominal Urology/gynaecology ENT/plastic Neurosurgery
SATISFY-SOS odds ratio (95% CI)
SATISFY-SOS MI odds ratio (95% CI)
Conditional NSQIP odds ratio (95% CI)
(n¼10 562)
P-value
(n¼11 986)
P-value
(n¼1590)
P-value
2.04 4.02 0.62 0.99 1.15 1.00
(1.78e2.35) (3.58e4.53) (0.53e0.73) (0.86e1.14) (1.03e1.28) (1.00e1.01)
<0.001 <0.001 <0.001 0.901 0.013 0.313
2.03 4.06 0.61 1.01 1.12 1.00
(1.78e2.32) (3.63e4.54) (0.53e0.71) (0.88e1.15) (1.01e1.25) (1.00e1.01)
<0.001 <0.001 <0.001 0.887 0.033 0.147
1.95 3.48 0.65 1.08 1.40 1.00
(1.26e3.04) (2.46e4.92) (0.42e1.02) (0.73e1.59) (1.02e1.91) (0.98e1.02)
0.003 <0.001 0.06 0.708 0.037 0.884
1.13 1.65 1.29 0.97 1.42 1.00
(0.88e1.45) (1.27e2.15) (0.85e1.97) (0.97e0.98) (1.27e1.60) (0.99e1.00)
0.296 <0.001 0.293 <0.001 <0.001 <0.001
1.10 1.56 1.22 0.97 1.43 0.99
(0.86e1.40) (1.21e2.01) (0.82e1.82) (0.97e0.98) (1.28e1.60) (0.99e1.00)
0.444 <0.001 0.327 <0.001 <0.001 <0.001
0.60 0.75 0.36 0.98 2.49 0.99
(0.24e1.51) (0.29e1.93) (0.09e1.39) (0.96e0.99) (1.82e3.42) (0.99e1.00)
0.279 0.552 0.138 <0.001 <0.001 0.057
1.61 1.16 1.54 1.24 0.98 1.03 1.34
(1.30e1.99) (0.86e1.55) (1.13e2.11) (0.97e1.59) (0.78e1.23) (0.80e1.32) (1.03e1.74)
<0.001 0.313 0.006 0.089 0.944 0.893 0.021
1.64 1.27 1.59 1.24 1.02 1.07 1.36
(1.34e2.01) (0.96e1.68) (1.18e2.15) (0.98e1.58) (0.82e1.27) (0.84e1.36) (1.06e1.74)
<0.001 0.093 0.002 0.078 0.857 0.564 0.017
1.69 1.49 1.49 1.15 0.93
d (0.92e3.11) (0.86e2.58) (0.96e2.3) (0.73e1.80) (0.44e1.98) d
d 0.092 0.153 0.073 0.555 0.849 d
and hyperalgesia.37,38 In the perioperative setting, medical complications may further exacerbate this inflammation, leading to persistent pain after surgery. The finding that no single complication appeared to be disproportionately responsible for the development of lingering pain could imply a systemic underlying mechanism, such as inflammatory or immune response, rather than a locally mediated effect. However, an association between severe acute pain on postoperative Day 1 and 30-day postoperative complications has been reported, questioning the directionality of the association.39 Therefore, future studies are required for understanding the precise mechanisms linking postoperative complications with persistent pain, and for finding biomarkers that can help identify high-risk patients in this context. This study has notable strengths. The data were collected prospectively using validated survey instruments from a large number of patients who underwent diverse elective surgical procedures. As such, there is a strong probability that these findings are generalisable. The credibility of the main finding is bolstered by the sensitivity analyses, including the incorporation of complications from the rigorous and established ACS-NSQIP registry. Furthermore, the association between medical complications and lingering post-surgical pain is consistent across different surgeries and medical complications affecting multiple organ systems, and after controlling for variables, such as age, sex, and the presence of pre-surgical pain. Similarly, the study has important limitations. First, with an observational design and retrospective analysis, causality cannot be inferred. Our findings should therefore be interpreted as hypothesis generating. Our pre-specified study protocol (Supplementary material) also slightly overestimated the number of available patients, because many entries in the initially obtained data set were for patients who underwent
anaesthesia for an excluded procedure (e.g. percutaneous intervention, colonoscopy, electroconvulsive therapy, etc.) or underwent numerous procedures. However, we were still able to include pain outcomes from a large and diverse surgical cohort. Second, our primary analysis utilised patient selfreport to assess complications, which may be subject to recall bias and inaccuracy because of patient misinterpretation, survey fatigue, or temporal ambiguity regarding perioperative events. A previous study using our patient-reported outcomes (PRO) method showed an excellent negative agreement, but only a low-to-moderate positive agreement in the detection of complications, when comparing PROs with complications documented in the electronic medical record.34 Whilst this finding might have revealed limitations in the PROs, it could also have reflected an advantage in that, with SATISFY-SOS, the patients reported complications that occurred both during hospitalisation and after discharge, whereas post-discharge complications are often not captured in the electronic medical record. We also considered that patients might be unsure about their medical diagnoses or complications, especially if they had more complicated hospital courses. To address this potential limitation, we conducted a sensitivity analysis using well-validated NSQIP complication data for patients enrolled in SATISFY-SOS, and reached similar findings. Third, the time frame of lingering pain outcome assessment in our study was generally within 1e3 months after surgery. Whilst it is highly probable that these outcomes are highly linked with the development of PPSP, the proposed cut-offs for which have ranged from 2 to 6 months after surgery,1 additional research is required for understanding the association between postoperative complications and outcomes beyond the 6-month postoperative time frame. Finally, our outcome of interest did not specify surgical site pain. The lack of pain location is somewhat mitigated by
6
-
Willingham et al.
Fig 2. Proportion with lingering post-surgical pain, and adjusted odds for lingering pain if a complication was also reported. Confidence intervals (CI) are Bonferroni corrected for eight regressions and include adjustment for baseline pain, age, sex, BMI, ASA physical status, preoperative opioid use, and mental component score. ENT, ear, nose, and throat; MIS, minimally invasive surgery; OR, odds ratio.
our inclusion of preoperative baseline pain (asked as exactly the same question) in our regression analysis. Of note, the non-specificity of this outcome may be appropriate, as inflammation from surgery and postoperative complications might globally sensitise patients to pain, including subsidiary pain from existing co-morbidities. In conclusion, this study adds to existing evidence by confirming in a large and diverse surgical cohort that lingering
post-surgical pain is common. Most importantly, our hypothesis that patients who experience postoperative medical complications are likely to be at much higher risk for pain that persists 1e3 months after surgery was supported. This finding, if replicated in future studies, suggests that clinicians should target patients who experience such complications for candidate interventions that might be effective in preventing postsurgical pain chronification. As long-term pain after surgery
Fig 3. Independent associations between complication categories and lingering post-surgical pain. Odds ratios are adjusted for baseline pain, age, sex, BMI, ASA physical status, mental component score, preoperative opioid use, and duration of follow-up. Odds ratios (95% confidence interval) from top to bottom are 1.44 (1.06e1.96), 1.53 (1.18e1.98), 1.56 (1.18e2.05), 1.60 (1.15e2.25), 1.66 (1.23e2.23), and 2.29 (1.74e3.01). GI, gastrointestinal.
Postoperative complications and lingering pain
is emerging as a major source of disability and persistent opioid use,2,9 these findings may have important implications for preventing these adverse consequences of surgery in highrisk patients.
Authors’ contributions Study conception/design: all authors. Funds acquisition: MSA. Data acquisition/analysis/interpretation: MW, GR, CC, ABA, AK, DH, MSA, SH. Statistical analyses: MW, GR. Drafting of manuscript: MW, GR. SH had full access to all the data in the study, and takes responsibility for the integrity of the data and the accuracy of the data analysis. He is the guarantor. All authors critically revised the manuscript for important intellectual content, and have approved the final version of the manuscript.
7.
8. 9.
10.
11.
12.
Acknowledgements The funding sources provided infrastructure and financial support, but had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication.
Declarations of interest
13.
14.
15.
The authors declare that they have no competing interests related to the publication or content of this publication. MSA is an editor of the British Journal of Anaesthesia. 16.
Funding Foundation for Barnes-Jewish Hospital (7937-77); Medical Student Anesthesia Research Fellowship grant from the Foundation for Anesthesia Education and Research to CC.
17.
18.
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.bja.2019.10.012.
19.
References 1. Werner MU, Kongsgaard UEI. Defining persistent postsurgical pain: is an update required? Br J Anaesth 2014; 113: 1e4 2. Clarke H, Soneji N, Ko DT, Yun L, Wijeysundera DN. Rates and risk factors for prolonged opioid use after major surgery: population based cohort study. BMJ 2014; 348: g1251 3. Kehlet H, Jensen TS, Woolf CJ. Persistent postsurgical pain: risk factors and prevention. Lancet 2006; 367: 1618e25 4. Kehlet H, Rathmell JP. Persistent postsurgical pain: the path forward through better design of clinical studies. Anesthesiology 2010; 112: 514e5 5. Haroutounian S, Nikolajsen L, Bendtsen TF, et al. Primary afferent input critical for maintaining spontaneous pain in peripheral neuropathy. Pain 2014; 155: 1272e9 6. Parsons B, Schaefer C, Mann R, et al. Economic and humanistic burden of post-trauma and post-surgical
20.
21.
22.
23.
-
7
neuropathic pain among adults in the United States. J Pain Res 2013; 6: 459e69 VanDenKerkhof EG, Hopman WM, Reitsma ML, et al. Chronic pain, healthcare utilization, and quality of life following gastrointestinal surgery. Can J Anaesth 2012; 59: 670e80 Macrae WA. Chronic pain after surgery. Br J Anaesth 2001; 87: 88e98 Brummett CM, Waljee JF, Goesling J, et al. New persistent opioid use after minor and major surgical procedures in US adults. JAMA Surg 2017; 152, e170504 Sun EC, Darnall BD, Baker LC, Mackey S. Incidence of and risk factors for chronic opioid use among opioid-naive patients in the postoperative period. JAMA Intern Med 2016; 176: 1286e93 Huang A, Azam A, Segal S, et al. Chronic postsurgical pain and persistent opioid use following surgery: the need for a transitional pain service. Pain Manag 2016; 6: 435e43 Soneji N, Clarke HA, Ko DT, Wijeysundera DN. Risks of developing persistent opioid use after major surgery. JAMA Surg 2016; 151: 1083e4 Wildgaard K, Ravn J, Kehlet H. Chronic post-thoracotomy pain: a critical review of pathogenic mechanisms and strategies for prevention. Eur J Cardiothorac Surg 2009; 36: 170e80 Aasvang EK, Luna IE, Kehlet H. Challenges in postdischarge function and recovery: the case of fast-track hip and knee arthroplasty. Br J Anaesth 2015; 115: 861e6 Clarke H. Transitional pain medicine: novel pharmacological treatments for the management of moderate to severe postsurgical pain. Expert Rev Clin Pharmacol 2016; 9: 345e9 Katz J, Jackson M, Kavanagh BP, Sandler AN. Acute pain after thoracic surgery predicts long-term post-thoracotomy pain. Clin J Pain 1996; 12: 50e5 Katz J, Seltzer Z. Transition from acute to chronic postsurgical pain: risk factors and protective factors. Expert Rev Neurother 2009; 9: 723e44 Katz J, Weinrib A, Fashler SR, et al. The Toronto General Hospital Transitional Pain Service: development and implementation of a multidisciplinary program to prevent chronic postsurgical pain. J Pain Res 2015; 8: 695e702 Nazemi S, Manaheji H, Noorbakhsh SM, et al. Inhibition of microglial activity alters spinal wide dynamic range neuron discharge and reduces microglial Toll-like receptor 4 expression in neuropathic rats. Clin Exp Pharmacol Physiol 2015; 42: 772e9 Yamamoto Y, Terayama R, Kishimoto N, et al. Activated microglia contribute to convergent nociceptive inputs to spinal dorsal horn neurons and the development of neuropathic pain. Neurochem Res 2015; 40: 1000e12 Zhou L-J, Yang T, Wei X, et al. Brain-derived neurotrophic factor contributes to spinal long-term potentiation and mechanical hypersensitivity by activation of spinal microglia in rat. Brain Behav Immun 2011; 25: 322e34 Duncan DJ, Yang Z, Hopkins PM, Steele DS, Harrison SM. TNF-alpha and IL-1beta increase Ca2þ leak from the sarcoplasmic reticulum and susceptibility to arrhythmia in rat ventricular myocytes. Cell Calcium 2010; 47: 378e86 Hotchkiss RS, Monneret G, Payen D. Immunosuppression in sepsis: a novel understanding of the disorder and a new therapeutic approach. Lancet Infect Dis 2013; 13: 260e8
8
-
Willingham et al.
24. Liu YY, Chiang CH, Chuang CH, Liu SL, Jheng YH, Ryu JH. Spillover of cytokines and reactive oxygen species in ventilator-induced lung injury associated with inflammation and apoptosis in distal organs. Respir Care 2014; 59: 1422e32 25. Singbartl K, Joannidis M. Short-term effects of acute kidney injury. Crit Care Clin 2015; 31: 751e62 26. Christiansen SC, Naess IA, Cannegieter SC, Hammerstrom J, Rosendaal FR, Reitsma PH. Inflammatory cytokines as risk factors for a first venous thrombosis: a prospective population-based study. PLoS Med 2006; 3: e334 27. Ghaferi AA, Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery. N Engl J Med 2009; 361: 1368e75 28. Costa MA, Trentini CA, Schafranski MD, Pipino O, Gomes RZ, Reis ES. Factors associated with the development of chronic post-sternotomy pain: a case-control study. Braz J Cardiovasc Surg 2015; 30: 552e6 29. Theunissen M, Peters ML, Schepers J, et al. Recovery 3 and 12 months after hysterectomy: epidemiology and predictors of chronic pain, physical functioning, and global surgical recovery. Medicine (Baltimore) 2016; 95, e3980 30. Helsten DL, Ben Abdallah A, Avidan MS, et al. Methodologic considerations for collecting patient-reported outcomes from unselected surgical patients. Anesthesiology 2016; 125: 495e504 31. von Elm E, Altman DG, Egger M, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ 2007; 335: 806e8
32. Fritz BA, Escallier KE, Ben Abdallah A, et al. Convergent validity of three methods for measuring postoperative complications. Anesthesiology 2016; 124: 1265e76 33. Hall BL, Hamilton BH, Richards K, Bilimoria KY, Cohen ME, Ko CY. Does surgical quality improve in the American College of Surgeons national surgical quality improvement program: an evaluation of all participating hospitals. Ann Surg 2009; 250: 363e76 34. Cohen ME, Liu Y, Ko CY, Hall BL. Improved surgical outcomes for ACS NSQIP hospitals over time: evaluation of hospital cohorts with up to 8 years of participation. Ann Surg 2016; 263: 267e73 35. Khuri SF. The NSQIP: a new frontier in surgery. Surgery 2005; 138: 837e43 36. Fragiadakis GK, Gaudilliere B, Ganio EA, et al. Patientspecific immune states before surgery are strong correlates of surgical recovery. Anesthesiology 2015; 123: 1241e55 37. Rowe RK, Ellis GI, Harrison JL, et al. Diffuse traumatic brain injury induces prolonged immune dysregulation and potentiates hyperalgesia following a peripheral immune challenge. Mol Pain 2016; 12: 1e12 38. Bugada D, Lavand’homme P, Ambrosoli AL, et al. Effect of preoperative inflammatory status and comorbidities on pain resolution and persistent postsurgical pain after inguinal hernia repair. Mediators Inflamm 2016; 2016: 5830347 39. van Boekel RLM, Warle MC, Nielen RGC, et al. Relationship between postoperative pain and overall 30-day complications in a broad surgical population: an observational study. Ann Surg 2019; 269: 856e65 Handling editor: Lesley Colvin